Perhaps a data point / leverage. The week -AFTER- the IETF, was the Yokohama International Quilt week. Same venue, same hotel. It’s been scheduled for over a year now. Many groups/tours are SOLD OUT, in planning to attend this event, plus side trips the week before and after. This event is larger than the IETF. …and it was recently canceled by the organizer… Perhaps (maybe) a number of the reserved rooms are tied to that event and they have not cleaned up after the effects of the cancelation. manning bmanning@xxxxxxxxxxx PO Box 6151 Playa del Rey, CA 90296 310.322.8102 On 17August2015Monday, at 10:18, Adam Roach <adam@xxxxxxxxxxx> wrote: > On 8/17/15 11:48, Ole Jacobsen wrote: >> Speaking from personal experience, I have always found Internet access >> in Japanese hotels to be quite excellent even without these upgrades >> by our NOC team. > > I suspect that the historically destroyed Internet connections in many of the overflow hotels -- and the Maastrict hotel for that matter -- are perfectly adequate for a normal mix of guests. I find it difficult to believe that you could accurately judge what a hotel's performance would be without a load similar to what IETF attendees typically bring with them. > > To be clear, issues rise above those of simple bandwidth saturation. Most commonly, I've seen things that I suspect are DHCP pool exhaustion (with results ranging from issuing duplicate addresses (!) to simply being unable to get an address) and NAT port exhaustion (leading to the inability to make or maintain connections). We bring a unique set of stresses to an infrastructure that are way outside the normal envelope. > > /a >